The objectives of this project are to identify genes and gene networks that play a role in the development of obesity-induced type 2 diabetes. In the previous grant period, we carried out the first-ever study that surveyed pancreatic islet gene expression in an intercross. The intercross was derived from mouse strains resistant (C57BL/6 or B6) or susceptible (BTBR) to diabetes induced by the leptinob mutation. In addition to islets, we surveyed gene expression in five other tissues (liver, adipose, muscle, hypothalamus, and kidney). All mice were genotyped for 5,000 single nucleotide polymorphisms between B6 and BTBR. We surveyed >100 clinical traits in the F2 sample, as well as hundreds of micro-RNAs (miRNAs) in liver. Our studies have yielded an unprecedented dataset, allowing us to develop causal regulatory networks that predict the interaction among genes, intermediate phenotypes such as miRNAs and mRNAs, and downstream clinical traits. Specifically these studies have allowed us to: 1) identify expression quantitative trait loci (eQTLs); 2) identify hotspots, or large sets of transcripts that appear to be co-regulated by a single locus; 3) use genetics as an anchor to develop directional eQTL network models; and 4) integrate eQTL network models with key physiological phenotypes related to obesity and diabetes. Taken together, the genetic architecture (eQTL and hotspots) and reconstructed networks have identified genomic regions and genes that may contain regulatory elements that underlie the physiology of obesity-dependent type 2 diabetes. We have used our methods for network construction to identify genes that are closely associated with clinical phenotypes. The specific aims of this proposal are to: 1) Identify and test regulatory genomic loci that govern pancreatic islet cell cycle genes essential for -cell replication. Refine and experimentally test models that predict the causal relationship between the regulatory loci and islet cell cycle genes responsible for the proliferative capacity of -cells; 2) Identify and test the regulatory genomic loci governing hepatic steatosis. Construct and refine network models that relate co-regulated hepatic genes with accumulation of liver triglycerides; 3) Utilize the heritability of miRNA abundance to predict mRNA targets and relate miRNAs, mRNAs and clinical phenotypes with testable models. Successful completion of Specific Aims 1 - 3 is critically dependent on the development of novel statistical and computational methods. These will include methods to: a) resolve genetic intervals to which multiple physiological and expression traits map; b) extend our network reconstruction approach to infer dependence among traits and incorporate prior information; and c) identify differentially correlated gene sets.